This paper is based on the theory of expectiles model and the definition of CARE model specifications, and it constructs IG-CARE and AS-CARE models in order to make an expectile-based VaR measure model that could estimate the interest rate risk of commercial bank. The results show that under the significant level of 5%, in which the fitting effect of IG-CARE model is not significant, the AS-CARE model has a better fitting effect, more stable estimated parameters, and remarkable leverage effect. Therefore, the AS-CARE model could be used as an effective measurement tool of the interest rate risk of commercial bank in China, and provide a theoretical basis and reference model for interest rate risk management.